Proc. of the 16 th Int. Conference on Digital Audio Effects (DAFx-13), Maynooth, Ireland, September 2-4, 2013 DAFX-1 BIT BENDING: AN INTRODUCTION Kurt James Werner, Mayank Sanganeria, Center for Computer Research in Music and Acoustics, Stanford University, Stanford Center for Computer Research in Music and Acoustics, Stanford University, Stanford California, USA California, USA kwerner@ccrma.stanford.edu mayank@ccrma.stanford.edu ABSTRACT We introduce the technique of "Bit Bending," a particularly fer- tile technique for circuit bending which involves short circuits and manipulations upon digital serial information. We present a justification for computer modeling of circuit-bent instruments, with deference to the movement's aversion to "theory-true" de- sign and associations with chance discovery [1]. To facilitate software modeling of Bit Bending, we also present a software library for modeling certain classes of digital integrated circuits. A synthesis architecture case study (frequency modulation via numerically controlled oscillators) demonstrates software model- ing of Bit Bending in action. 1. INTRODUCTION Before discussing Bit Bending, we will briefly review the context of circuit bending and the motivation for modeling musical in- struments and audio effects in software, as well as a justification for computer modeling of circuit-bent instruments in particular. 1.1. History of circuit bending Circuit bending is the process of creatively modifying or aug- menting sound-producing electronic devices [2], typically toy instruments, guitar effect pedals, sound-producing children’s toys, etc. Circuit bending is primarily associated with experimental, noise, do it yourself (DIY), hacker, chiptunes, and multimedia art scenes. Recently, the practice has found a foothold in the main- stream and occupies an increasingly important musical and cul- tural niche [3]. High profile artists like Radiohead, the Flaming Lips, and Björk increasingly experiment with and utilize circuit- bent instruments [3]. Information about circuit bending is shared freely online [4, 5], and DIY project communities (including cir- cuit bending interest groups) thrive in various major metropolitan areas through workshops, festivals, and meetups [6]. Interest in circuit-bent instruments stems from the rich and com- plex sonic palette they create, the nostalgic associations conjured by modifying old devices, the repurposing of old technology, and a preoccupation with the emancipation of noise in musical prac- tice [7]. In the context of “clock bends” or the use of binary counters to manipulate data streams (a special case of Bit Bend- ing that will be discussed in this paper), circuit bending brings its own non-linear flair to a tradition of multi-scale sound design and composition [8, 9]. Figure 1: Circuit-bent Speak & Spell with MIDI control, Kurt James Werner, 2007 Though the practice began in the 1960s [10] (and traces roots to Leon Theremin's experiments with radio tubes in the 1920s [11]), it is still understudied. An example of a circuit-bent device is shown in Figure 1. 1.2. Overview of modeling techniques Models are used to predict the behavior of a system, and model- building is a very useful way of understanding the world [12: 2]. It is often desirable to model the behavior of musical instruments and effects in software [12: 3]. There are numerous approaches to modeling, which are broadly categorized as non-physical signal models (spectral models, etc.) [12: 4], and physical signal models (which leverage “an explicit representation of the relevant physical state of the sound source”) [12: 6]. One could also categorize modeling approaches into “top-down” (or “feature synthesis”) models and “bottom-up” (or “system synthesis”) models. A top-down approach would identi- fy high-level features of a system that are important / perceptible / desirable and then try to recreate them in a model. A bottom-up approach would seek to understand and model each small foun- dational component, inter-connecting them appropriately, assum- ing that if each component is accurately modeled, the high-level features will occur as a by-product. Proc. of the 16th Int. Conference on Digital Audio Effects (DAFx-13), Maynooth, Ireland, September 2-5, 2013